It's a bit of a combination of both the science being not able to test the whole DNA strand, and that's why...because 99.9% of all our DNA is the same, it's only that 0.1% that differentiates us all. And on the strand they look at—and you've probably had this presented by other speakers in earlier days—loci, in terms of how many.
When the system first came out, they were looking at six, and they said, “It's impossible, you won't find another of six”, and then there was a false match in the U.K. on six. Then they went to nine, and they said, “With nine, you're not going to have any problems”. And again, the Arizona thing I mentioned, there are a lot of false matches in nine. Now they're saying the same about 10 and 11. I know at the last Promega Conference this past October, where DNA scientists all around the world get together, they said yes, they have seen false matches at 10 and 11. And that's why there are all these challenges of state databases now in the U.S., because they want to see what the chances are that 13 is going to have false matches.
So until the whole strand can be tested, and that's when it will be absolutely foolproof, there's always that chance. It's just like subway cars. We all have 20 cars, and if you're only testing four, there might be a chance unless you go to five. And that's why I say the bigger the database, the more chances of false matches.